dspoon@fenix.Atlanta.NCR.COM (David B. Witherspoon) (03/01/88)
Recently we added several hundred feet of cable to our Token Ring LAN. Since then, we have been noticing MANY problems with the server software (vendor intentionally not mentioned) which seem to coincide with ring errors. My question is: What is the general "psychology" of a TR LAN in terms of ring errors recovery? Unfortunately, we were not monitoring the ring status continually until we started having problems, so I don't know if the ring errors have increased since the new cabling was installed (which would seem to indicate a H/W bug) or not (which would seem to indicate a S/W bug). The ring occasionally will suffer a "Ring Error", which is always followed by a "Ring Recovered", the time between the two being around 20 seconds. The server software appears to be intolerant of such conditions and various problems ensue, usually requiring rebooting of the server(s). The questions restated is: Is a TR LAN designed to suffer and deal with (recover from) various kinds of ring errors, from which the S/W should be able to recover? If there is a good source of info on T/R LAN debugging and/or any popular S/W tools, I'd appreciate that info as well. Thanks in advance!! -------------------------------David Witherspoon------------------------------- D.Witherspoon@Atlanta.NCR.COM | "Facts all come with points of view NCR Sys Engineering-Retail/Atlanta | Facts don't do what I want them to" MY OPINIONS...ALL MINE!!! | - Talking Heads
kevin@vger.NBI.COM (Kevin Brooks) (03/05/88)
In article <302@fenix.Atlanta.NCR.COM> dspoon@fenix.Atlanta.NCR.COM (David B. Witherspoon) writes: > >The questions restated is: > > Is a TR LAN designed to suffer and deal with (recover from) various > kinds of ring errors, from which the S/W should be able to recover? > >If there is a good source of info on T/R LAN debugging and/or any popular S/W >tools, I'd appreciate that info as well. Thanks in advance!! > >-------------------------------David Witherspoon------------------------------- >D.Witherspoon@Atlanta.NCR.COM | "Facts all come with points of view >NCR Sys Engineering-Retail/Atlanta | Facts don't do what I want them to" >MY OPINIONS...ALL MINE!!! | - Talking Heads The token ring has two levels of communication. The first being the DLC (data link control) which provides - Connnectionless communication between devices providing no guarantee of delivery. These are used for such things as broadcast and datagram messages. The second is the connection oriented services useing the LLC (logical link level) which provides - A reliable transport mechanism in which acknowledgment and retransmission are possible. As to answer your question on TR lan debugging try the IBM token ring problem and determination guide. IBM pn GA27-3361. But to tell you in short what it recomends is to use your status information provided from your observer terminal which runs a net manager or token ring diagnostic program. With that information you will know if you are encountering soft or hard errors and if a device is beaconing and the NAUN (nearest active upstream neighbor) after which you will have a fault doamin isolated. At that time you would disconnect the fault domain and see if the error persist, etc..... Hope this helps -- Kevin Brooks Usenet: ...{pyramid!isieng}hao|nbires}!vger!kevin